Deep learning-assisted cone-beam computed tomography volumetric assessment of mandibular anterior alveolar bone changes with and without periodontally accelerated osteogenic orthodontics during skeletal Class III orthodontic−orthognathic treatment

Introduction Patients with skeletal Class III malocclusion often have thin anterior alveolar bone and an increased periodontal risk. This study evaluated 3-dimensional (3D) alveolar bone volume changes in patients treated with or without periodontally accelerated osteogenic orthodontics (PAOO) using a deep learning-assisted cone-beam computed tomography (CBCT) approach. Methods A total of 42 adults undergoing orthodontic−orthognathic treatment were divided into a PAOO group (group S, n=21) and a matched non-PAOO control group (group NS, n=21). CBCT scans were obtained at baseline (T1) and posttreatment (T3), with a 6-month post-PAOO scan (T2) for group S. Labial, lingual, and total alveolar bone volumes were quantified using a deep learning-assisted automated segmentation tool. Linear mixed-effects models were used to analyze longitudinal changes and treatment effects, adjusting for baseline root length. Results In group S, labial alveolar bone volume increased from T1 (42.94 ± 23.15 mm 3 ) to T3 (84.15 ± 36.82 mm 3 ; P <0.001), contributing to an increase in total alveolar bone volume (139.92 ± 86.42 mm 3 at T3 vs 114.20 ± 66.86 mm 3 at T1; P <0.001). Group NS showed no change in labial bone volume ( P = 0.846) but a reduction in total bone volume at T3 (93.76 ± 67.24 mm 3 ; P <0.001), primarily driven by lingual loss. Group S had greater labial and total alveolar bone volumes than group NS at T3 ( P <0.001). Conclusions PAOO was associated with increased and maintained labial alveolar bone volume, which may help preserve total alveolar support during combined treatment. Deep learning-assisted CBCT segmentation can serve as an efficient volumetric evaluation tool.